Previous   |    Next

doi:10.2204/iodp.proc.340.204.2016

Methods

Core sites

Sites U1394, U1395, and U1396 were drilled during March 2012. Site U1394 is located ~10 km southwest of Montserrat, and Site U1395 is located ~25 km southwest of Montserrat. Sites U1394 and U1395 are located within the Boulliant-Montserrat Graben, which funnels turbidites originating from southeast of Montserrat. Two holes were drilled at each of Sites U1394 and U1395. Hole U1394A recovered 24.17 m of core with ~41% disturbed core, and Hole U1394B recovered 137.42 m of core with an estimated core disturbance of ~29%. Hole U1394A had poor core recovery below 35 m because of the rotary drilling technique used below this depth. Hole U1394B has a gap between 15 and 60 mbsf because of the rotary drilling technique; a subsequent change to advanced piston coring produced much better core recovery. Hole U1395A recovered 124.29 m of core with 7%–13% estimated core disturbance, and Hole U1395B recovered 127.52 m of core with 12%–16% core disturbance.

Site U1396 is located ~35 km southeast of Montserrat on a topographic high. Three holes were drilled at Site U1396. Hole U1396A recovered 140.51 m of core with <1% estimated core disturbance, and Hole U1396C recovered 145.92 m of core with 3% estimated core disturbance. The interval between 5 and 15 mbsf was deformed in Hole U1396A during drilling; therefore, a short (10 m) third hole (U1396B) was drilled to recover this 5–15 m interval. Holes U1394B, U1395B, and U1396C were the most complete; thus, initial work has primarily been conducted on these holes.

Core logging

Initial core logging was done during Expedition 340 on the R/V JOIDES Resolution, including detailed graphical logs (at a scale of 1–10 cm) and analysis of grain compositions. This ship-based work is captured on the visual core descriptions (VCDs) (see “Core descriptions”). Further core logging of key core sections was completed at the IODP Gulf Coast Repository (GCR) in College Station, Texas (USA), after the expedition, together with more analyses of grain composition (“Appendix A”).

Identification of coring artifacts—especially sand sucked in during piston coring

It is important to distinguish in situ deposits from the artifacts of coring. This issue is analyzed in more detail by Jutzeler et al. (2014). In particular, advanced piston coring can suck in layers of massive sand, which could be mistaken for turbidites. Jutzeler et al. (2014) outline a method for identifying which sand layers are an artifact of piston core suck-in. For instance, these sand layers tend to terminate at the bottom of the individual sections of IODP piston core. Sand layers that may have been sucked in are shown by shaded boxes on the logs in “Appendix A” and “Appendix B.”

Core sampling

Cores were stored at the GCR. Samples were taken from hemipelagic mud for oxygen isotope analysis and also from every event deposit to determine type and emplacement mechanism of each event found within the cores. Sample volumes varied between >1 and 10 cm³ depending on abundance and thickness of the unit; care was taken to avoid disturbed sections of the core. Sample processing and analysis was conducted at the National Oceanography Centre, Southampton (NOCS, United Kingdom), and Plymouth University (United Kingdom). All samples were dried in a cool oven at 40°C for 24 h prior to removal of fines by wet sieving over a 63 µm sieve. Both fine and coarse fractions were collected and dried in a cool oven for 24 h. Some sample for grain-size analysis was kept aside and not sieved.

Grain sizes

Grain size distributions were used to distinguish between submarine turbidity current and fallout deposits, as outlined below. Laser diffraction was used to measure grain size distributions using a Malvern (Mastersizer 2000) particle size analyzer. This instrument has a range of 0.2–2000 µm. A volume of 25 mL of reverse osmosis water with 0.05% sodium hexametaphosphate dispersant was added to 1 cm³ of unsieved sample and left overnight on a shaking table. Samples were analyzed three times using the Mastersizer, and accuracy was monitored using standard size particles (32 and 125 µm).

Componentry—composition of layers

The composition of layers provides key information on their origin (“Appendix B”). Initial component analyses were completed on the vessel during Expedition 340, dividing grains into one of seven classifications:

1. Vesicular pumice,
2. Scoria clasts,
3. Nonvesiculated lava clasts,
4. Vesiculated lava clasts,
5. Volcaniclastic clasts,
6. Sedimentary clasts including bioclasts, or
7. Other clasts (e.g., hydrothermal clasts).

Using smear slides, abundance estimates were made using an area chart. Data are shown in “Appendix A.”

More detailed componentry analysis was conducted at NOCS using >63 ?m material from all volcanic-rich units and some bioclastic-rich units from Holes U1394B, U1395B, U1396A, and U1396C. For each sample, at least 300 grains were point-counted using a field counting method, where all grains within the microscope field of view are counted. Grains were divided into one of six classifications:

1. Vesicular pumice clasts,
2. Nonvesicular lava,
3. Altered lithic clasts,
4. Crystal and glass fragments,
5. Mafic scoria clasts, or
6. Bioclasts.

This follows the classification of Le Friant et al. (2 s008) and Cassidy et al. (2014b). Data are shown in “Appendix A” and Table 204_ST03.CSV.

Classification of deposit types in cores

Five main types of deposit are defined: hemipelagic mud, bioclastic turbidites, mixed turbidites, volcaniclastic turbidites, and tephra fall layers.

Hemipelagic mud is easily identifiable, as it has relatively high mud content with a brown-yellow color and interspersed foraminifers. However, in some cases hemipelagic mud can be difficult to distinguish from a bioclast-rich turbidite, as both may comprise muddy sand rich in bioclasts. This situation occurs when the hemipelagic mud has a higher sand content, perhaps due to contamination by material winnowed by ocean currents.

It can be problematic to distinguish turbidity current deposits (turbidites) from tephra fall layers. We refer the reader to Cassidy et al. (2014b) for a fuller discussion of how to distinguish fallout deposits from turbidites. Here we define the deposit types according to the following consistent criteria:

• Bioclastic turbidites: >70% bioclasts.
• Mixed turbidites: 30%–70% bioclasts.
• Volcaniclastic turbidites: <30% bioclasts and a Folk and Ward (1957) sorting coefficient >0.5 (phi).
• Tephra fall: <30% bioclasts, Folk and Ward (1957) sorting coefficient <0.5 (phi), and <20 cm thick.

Pb isotopes

Montserrat is situated in an active island arc. Guadeloupe is a volcanic island nearby both Montserrat and the IODP sites; therefore, visible tephra layers from Guadeloupe are likely to feature within the IODP holes. Using lead isotopes, deposits from Montserrat and Guadeloupe can be distinguished, as volcanic rocks from Guadeloupe are relatively enriched in radiogenic lead (Cassidy et al., 2012).

Tephra from Hole U1396C was analyzed for lead isotopes. For each sample, 200 mg of pumice or scoria clasts was picked out and leached in 4 mL of 6 M HCl at 140°C for 1–2 h in sealed Teflon vials. Samples were then left on a hot plate at 130°C for 24 h to dissolve in HF-HNO₃. Samples were left to dry, and then a further 0.5 mL of concentrated HCL and 0.5 mL of concentrated HNO₃ were added, evaporating until dry between each addition. To the remaining residue, 1.5 mL of HBr was added and heated for 1 h. A supernatant for column chemistry was then produced by centrifuging samples for 5 min. Isolation of Pb from the matrix was achieved using AG1-X8 200–400 mesh anion exchange resin. Blanks contained <50 pg of Pb, and the Pb standard NBS 981 was used. Isotope analyses were conducted on a VG Sector 54 thermal ionization mass spectrometer and multiple collector inductively coupled plasma mass–spectrometer (MC-ICP-MS) (Neptune and GV Iso Probe) at NOCS. The double-spike method was used to correct instrumental bias (Ishizuka et al., 2003). Sample preparation and analyses methods follow those of Cassidy et al. (2012).

Dating

Oxygen and carbon isotope stratigraphies

The global oxygen isotope curve of seawater fluctuates through time from higher to lower values (Lisiecki and Raymo, 2005). The global oxygen isotope curve is divided into dated marine isotope stages (MIS) representing periods of either higher or lower isotopic values of the global ocean. Marine isotope stages can be identified at Sites U1394, U1395, and U1396 by comparing their respective oxygen isotope records with the global record. Holes U1394B, U1395B, and U1396C were most complete; thus, oxygen isotope stratigraphy was conducted primarily on these holes. However, the upper part of Hole U1394A was more complete than Hole U1394B; thus, the upper 6 m of Hole U1394A was sampled. Oxygen and carbon isotope analyses were carried out for deposits comprising the upper ~250 ky period of each site, corresponding to the upper 6 m of Hole U1394A and 92–166 m of Hole U1394B, the upper 44 m of Hole U1395B, and the upper 7 m of Hole U1396C. For isotope analyses, hemipelagic intervals were sampled every 7 cm for Holes U1394A, U1394B, and U1395B. For Hole U1396C, hemipelagic intervals were sampled every 5 cm. Oxygen isotope stratigraphy was used in combination with biostratigraphy and AMS dating to improve identification of marine isotope stages and sedimentary hiatuses at these sites.

Samples were analyzed at Plymouth University. Twenty Globigerinoides ruber specimens between 250 and 355 µm in size were picked from each hemipelagic sample and reacted with 100% phosphoric acid at 90°C for 2 h. The resulting CO₂ was analyzed using an Isoprime Instruments continuous flow mass spectrometer with a Gilson multiflow carbonate autosampler. Oxygen and carbon isotope values (δ¹⁸O and δ¹³C) are given as per mil (‰) deviations in the isotope ratios (¹⁸O/¹⁶O and ¹³C/¹²C) calibrated against Vienna Peedee belemnite (VPDB) using internal standards NBS-19, IAEA-CO-8, and IAEA-CO-9. Five NBS-19 standards were also evenly distributed throughout the individual isotope runs to correct for daily drift. The mean standard deviation on replicate analyses was 0.17‰ for δ¹⁸O and 0.19‰ for δ¹³C.

AMS radiocarbon dating

Four samples from the upper 4 m of Hole U1395B (aged < 57 ka) were AMS dated as part of this study. Site U1395 is located near the drill site of Core JR123-12V, described by Trofimovs et al. (2013) (Figs. F1, F2). AMS samples selected were beneath two of the largest turbidites in the uppermost 4 m of Hole U1395B (Fig. F2). The upper turbidite is a normally graded bioclastic turbidite 107 cm thick; the lower turbidite is a normally graded mixed turbidite 116 cm thick. Two samples were taken beneath each turbidite. For further details see Table T1 and Figure F2.

Approximately 1000 pristine (not reworked) tests of Globigerinoides ruber >150 µm in size were picked (~17 mg) and sonically cleaned for each sample. Radiocarbon dates were measured at Scottish Universities Environmental Research Council (SUERC) using their in-house protocol (see Trofimovs et al., 2013). The AMS dates for the lower two analyses in Hole U1395B are close to the analytical limit and so are reported as an uncalibrated age of older than 45 ka.

Biostratigraphy

Offshore of Montserrat, planktonic foraminifer datum species do not follow the standard zonation of Wade et al. (2011) during the last 250 ky (Wall-Palmer et al., 2014) and were not used for stratigraphy. However, the distribution of Globorotalia menardii was used, following the zonation previously published offshore of Montserrat by Le Friant et al. (2008).

Calcareous nannofossils in the <63 µm fraction of hemipelagic samples were analyzed using scanning electron microscopy (SEM). Dry pieces of sediment were adhered to metal stubs using silver paint then sputter-coated with gold. The zonation of Kameo and Bralower (2000) for the Caribbean Sea was used to determine the nannofossil stratigraphy for the three sites.

Paleomagnetism

Paleomagnetism can be used to date Holes U1394B, U1395B, and U1396C. During the deposition of volcanic and sedimentary rocks, magnetic minerals align to Earth’s magnetic field. When the rocks are compacted, the magnetic minerals become fixed, thus preserving the direction of Earth’s magnetic field within the rock record. Periodically, Earth’s polarity reverses; the ages of these magnetic reversals are well constrained. Pole reversals can be found within Holes U1394B, U1395B, and U1396C by analyzing the natural remnant magnetization (NRM) of the cores.

Data were collected during Expedition 340 (Hatfield, 2015; also see the “Methods” chapter [Expedition 340 Scientists, 2013]). Archive-half core sections were analyzed every 2.5 cm before and after alternating field demagnetization in a peak field of 20 mT and an additional step of 10 mT, if time and core flow allowed. NRM measurements were also made over a 15 cm long interval before and after each core section and were used to monitor background magnetic moment and allowed deconvolution of the response of the pickup coils. With no onboard deconvolution, measurements from the top and bottom 10 cm of each section were omitted from analysis, as these regions are most susceptible to volumetric edge effects associated with instrument response functions. Analyses were conducted using a 2G Enterprises model 760R superconducting rock magnetometer (SRM) with superconducting quantum interference devices (SQUIDs) and an in-line automated alternating field demagnetizer. Pickup coil response functions for x-, y-, and z-axes were 6.1, 6.2, and 9.9 cm, respectively. The measured area at each interval is then integrated over ~100 cm³ (Richter et al., 2007). Measurement of an empty tray followed by background drift correction allowed estimation of the ambient noise level at ~2 × 10^–6A/m. For further details see the “Methods” chapter (Expedition 340 Scientists, 2013). Revised reversal ages from Ogg et al. (2012) were used.

Top of page   |    Previous   |    Next